fix(qa): scope mock image prompts to latest turn

extensions/qa-lab/src/providers/mock-openai/server.test.ts

@@ -3129,6 +3129,53 @@ describe("qa mock openai server", () => {
     expect(text).not.toBe("ui bridge armed");
   });
 
+  it("keeps stale image prompts from overriding later marker turns", async () => {
+    const server = await startMockServer();
+
+    const response = await fetch(`${server.baseUrl}/v1/responses`, {
+      method: "POST",
+      headers: { "content-type": "application/json" },
+      body: JSON.stringify({
+        stream: false,
+        model: "mock-openai/gpt-5.5",
+        input: [
+          {
+            role: "user",
+            content: [
+              {
+                type: "input_text",
+                text: "Image understanding check: describe the top and bottom colors.",
+              },
+              {
+                type: "input_image",
+                source: {
+                  type: "base64",
+                  mime_type: "image/png",
+                  data: QA_IMAGE_PNG_BASE64,
+                },
+              },
+            ],
+          },
+          {
+            role: "assistant",
+            content: [
+              {
+                type: "output_text",
+                text: "Protocol note: the attached image is split horizontally, with red on top and blue on the bottom.",
+              },
+            ],
+          },
+          makeUserInput("Marker exact marker: `fresh-marker-ok`"),
+        ],
+      }),
+    });
+    expect(response.status).toBe(200);
+    const payload = (await response.json()) as {
+      output?: Array<{ content?: Array<{ text?: string }> }>;
+    };
+    expect(payload.output?.[0]?.content?.[0]?.text).toBe("fresh-marker-ok");
+  });
+
   it("handles deeply nested image input shapes without recursive traversal failure", async () => {
     const server = await startQaMockOpenAiServer({
       host: "127.0.0.1",

extensions/qa-lab/src/providers/mock-openai/server.ts

@@ -555,6 +555,35 @@ function countImageInputs(value: unknown): number {
   return count;
 }
 
+function extractLatestImageUserTurn(input: ResponsesInputItem[]) {
+  const latestUserIndex = findLastUserIndex(input);
+  if (latestUserIndex < 0) {
+    return { text: "", imageInputCount: 0 };
+  }
+
+  let startIndex = latestUserIndex;
+  while (
+    startIndex > 0 &&
+    input[startIndex - 1]?.role === "user" &&
+    Array.isArray(input[startIndex - 1]?.content)
+  ) {
+    startIndex -= 1;
+  }
+
+  const imageTurnItems = input.slice(startIndex, latestUserIndex + 1);
+  const imageInputCount = countImageInputs(imageTurnItems.map((item) => item.content));
+  if (imageInputCount === 0) {
+    return { text: "", imageInputCount: 0 };
+  }
+  return {
+    text: imageTurnItems
+      .map((item) => extractInputText(item.content as unknown[]))
+      .filter(Boolean)
+      .join("\n"),
+    imageInputCount,
+  };
+}
+
 function parseToolOutputJson(toolOutput: string): Record<string, unknown> | null {
   if (!toolOutput.trim()) {
     return null;
@@ -1009,7 +1038,7 @@ function buildAssistantText(
     extractExactMarkerDirective(prompt) ?? extractExactMarkerDirective(allInputText);
   const finishExactlyDirective =
     extractFinishExactlyDirective(prompt) ?? extractFinishExactlyDirective(allInputText);
-  const imageInputCount = countImageInputs(input);
+  const latestImageUserTurn = extractLatestImageUserTurn(input);
   const activeMemorySummary = extractActiveMemorySummary(allInputText);
   const snackPreference = extractSnackPreference(activeMemorySummary ?? memorySnippet);
   const sessionsSpawnError = extractToolErrorForNamedCall({
@@ -1036,10 +1065,16 @@ function buildAssistantText(
   if (isHeartbeatPrompt(prompt)) {
     return "HEARTBEAT_OK";
   }
-  if (/roundtrip image inspection check/i.test(allInputText) && imageInputCount > 0) {
+  if (
+    /roundtrip image inspection check/i.test(latestImageUserTurn.text) &&
+    latestImageUserTurn.imageInputCount > 0
+  ) {
     return "Protocol note: the generated attachment shows the same QA lighthouse scene from the previous step.";
   }
-  if (/image understanding check/i.test(allInputText) && imageInputCount > 0) {
+  if (
+    /image understanding check/i.test(latestImageUserTurn.text) &&
+    latestImageUserTurn.imageInputCount > 0
+  ) {
     return "Protocol note: the attached image is split horizontally, with red on top and blue on the bottom.";
   }
   if (/\bmarker\b/i.test(allInputText) && exactReplyDirective) {
@@ -1565,7 +1600,7 @@ async function buildResponsesPayload(
     extractExactReplyDirective(prompt) ?? extractExactReplyDirective(allInputText);
   const exactMarkerDirective =
     extractExactMarkerDirective(prompt) ?? extractExactMarkerDirective(allInputText);
-  const imageInputCount = countImageInputs(input);
+  const latestImageUserTurn = extractLatestImageUserTurn(input);
   const firstExactMarkerDirective = extractLabeledMarkerDirective(
     allInputText,
     "first exact marker",
@@ -1652,12 +1687,18 @@ async function buildResponsesPayload(
   if (/fanout worker beta/i.test(prompt)) {
     return buildAssistantEvents("BETA-OK");
   }
-  if (/roundtrip image inspection check/i.test(allInputText) && imageInputCount > 0) {
+  if (
+    /roundtrip image inspection check/i.test(latestImageUserTurn.text) &&
+    latestImageUserTurn.imageInputCount > 0
+  ) {
     return buildAssistantEvents(
       "Protocol note: the generated attachment shows the same QA lighthouse scene from the previous step.",
     );
   }
-  if (/image understanding check/i.test(allInputText) && imageInputCount > 0) {
+  if (
+    /image understanding check/i.test(latestImageUserTurn.text) &&
+    latestImageUserTurn.imageInputCount > 0
+  ) {
     return buildAssistantEvents(
       "Protocol note: the attached image is split horizontally, with red on top and blue on the bottom.",
     );